{ "_name_or_path": "pszemraj/beit-large-patch16-512-types-of-film-shots-vN", "add_fpn": false, "architectures": [ "BeitForImageClassification" ], "attention_probs_dropout_prob": 0.0, "auxiliary_channels": 256, "auxiliary_concat_input": false, "auxiliary_loss_weight": 0.4, "auxiliary_num_convs": 1, "drop_path_rate": 0.1, "finetuning_task": "image-classification", "hidden_act": "gelu", "hidden_dropout_prob": 0.0, "hidden_size": 1024, "id2label": { "0": "ambiguous", "1": "closeUp", "2": "detail", "3": "extremeLongShot", "4": "fullShot", "5": "longShot", "6": "mediumCloseUp", "7": "mediumShot" }, "image_size": 512, "initializer_range": 0.02, "intermediate_size": 4096, "label2id": { "ambiguous": "0", "closeUp": "1", "detail": "2", "extremeLongShot": "3", "fullShot": "4", "longShot": "5", "mediumCloseUp": "6", "mediumShot": "7" }, "layer_norm_eps": 1e-12, "layer_scale_init_value": 0.1, "model_type": "beit", "num_attention_heads": 16, "num_channels": 3, "num_hidden_layers": 24, "out_features": [ "stage24" ], "out_indices": [ 24 ], "patch_size": 16, "pool_scales": [ 1, 2, 3, 6 ], "problem_type": "single_label_classification", "reshape_hidden_states": true, "semantic_loss_ignore_index": 255, "stage_names": [ "stem", "stage1", "stage2", "stage3", "stage4", "stage5", "stage6", "stage7", "stage8", "stage9", "stage10", "stage11", "stage12", "stage13", "stage14", "stage15", "stage16", "stage17", "stage18", "stage19", "stage20", "stage21", "stage22", "stage23", "stage24" ], "torch_dtype": "float32", "transformers_version": "4.38.0.dev0", "use_absolute_position_embeddings": false, "use_auxiliary_head": true, "use_mask_token": false, "use_mean_pooling": true, "use_relative_position_bias": true, "use_shared_relative_position_bias": false, "vocab_size": 8192 }